[
https://issues.apache.org/jira/browse/HADOOP-15961?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16744885#comment-16744885
]
Steve Loughran commented on HADOOP-15961:
-----------------------------------------
OK.
Overall design looks good, just some of the details to tune.
# doesn't apply to trunk; HADOOP-14556 has broken it. Sorry..you'll need to
tweak it there and we'll have to go the other way on the backporting (which
will be appropriate)
# Given that progressible is just an interface there's no need to mock &
verify, just use a simple implementation
{code}
class ProgressCounter implements Progressible {
private long count;
public void progress() { count++; }
}
{code}
..then do some asserts on that count value.
style wise, there's a few changes for the indentation...ideally the patch
shouldn't be realigning things. That's pretty minor.
> S3A committers: make sure there's regular progress() calls
> ----------------------------------------------------------
>
> Key: HADOOP-15961
> URL: https://issues.apache.org/jira/browse/HADOOP-15961
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/s3
> Reporter: Steve Loughran
> Assignee: lqjacklee
> Priority: Minor
> Attachments: HADOOP-15961-001.patch
>
>
> MAPREDUCE-7164 highlights how inside job/task commit more context.progress()
> callbacks are needed, just for HDFS.
> the S3A committers should be reviewed similarly.
> At a glance:
> StagingCommitter.commitTaskInternal() is at risk if a task write upload
> enough data to the localfs that the upload takes longer than the timeout.
> it should call progress it every single file commits, or better: modify
> {{uploadFileToPendingCommit}} to take a Progressable for progress callbacks
> after every part upload.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
